Flawed CDC gun violence data concerns researchers

Public health researchers are alarmed by an analysis of U.S. Centers for Disease Control and Prevention (CDC) data showing that the number of Americans nonfatally injured by a firearm jumped by 37% between 2015 and 2016.

Led by FiveThirtyEight and The Trace, the analysis—outlined in an October 4, 2018 FiveThirtyEight article—found that the CDC flagged its own data as “unstable and potentially unreliable,” yet an agency spokesperson told reporters that CDC experts are “confident that the sampling and estimation methods are appropriate.”

However, FiveThrityEight and The Trace shared their review with more than a dozen public health researchers, many of whom said the inaccuracy of the CDC data is troubling and could skew the national-level understanding of gun violence.

Among the researchers who expressed concern was Harvard T.H. Chan School of Public Health’s David Hemenway, director of the Harvard Injury Control Research Center. “No one should trust the CDC’s nonfatal firearm injury point estimates,” Hemenway said. “Basically, the confidence intervals are enormous, so you have no idea about trends.”

Read the FiveThirtyEight article: The CDC Is Publishing Unreliable Data On Gun Injuries. People Are Using It Anyway.